154 



unlike those of Robertson, are acid in character, or at least in tendency, 

 and the lime is probably present in a different state of combination : 

 at all events, the boiling and evaporation of the soil, according to Method 

 III., with concentrated Hydrochloric and Nitric acids, yielded from twice 

 to six times as much lime as agitation with diluted Hydrochloric acid 

 in the cold, as prescribed by Method I. In the case of potash the differ- 

 ence was even more marked, Method III. extracting from six to eleven 

 times as much as Method I. It will be noticed that the difference between 

 the two methods was greatest, not only with regard to lime, but also in 

 respect to potash, in the third members of the two sets of soils, namely 

 Nos. 3 and 22, and it was least in the second, i.e., Nos. 2 and 21. 



If, for the reasons already given, Method I. is locally preferable to 

 Method IV., much more then is it to be preferred to Method III. 



In addition to the three George soils, most of the samples taken in 

 the Mossel Bay and Riversdale Divisions were extracted according to 

 Method III. No soils from the latter two divisions have been examined 

 by Method I., but they may be expected to yield results not widely differ- 

 ing from and rather below those of the Swellendam soils. The means 

 of comparing the methods in these areas are not very full, the only Swel- 

 lendam soils examined by Method I. being those of the Tradouw, in 

 which the average proportions of lime so obtained fall greatly below those 

 yielded by Method III. for the Riversdale and Mossel Bay soils. In 

 this respect, as also in regard to potash, the latter show results* very 

 similar to those obtained from the George soils by the same method, and 

 so the general conclusion may be drawn that the method yields results 

 too high to be sufficiently comparable with those attained by natural 

 methods, and hence that considerable diminutions will have to be made 

 in its figures if the chemical aspects of soil fertility are to be judged 

 by them. The proportions of lime and potash, for instance, yielded by 

 No*. 5 of the Mossel Bay Division soils, poor and acid as it is reputed to 

 be, would, if obtained by Method I., have gained for it the declaration 

 of being a soil adequately provided with these plant food elements, and 

 practical experience would have dissented from such a view. A similar 

 observation may be made regarding Nos. 16 and 17 of the Mossel Bay 

 soils, and Nos. 18 and 24 of those from Riversdale. 



The obvious reasons for discarding methods which were inherently 

 more energetic than that eventually adopted did not seem to apply to 

 Method II., and hence- it was given a somewhat extensive trial. It is 

 difficult to conceive why it should have yielded results frequently so 

 much higher than those of Method I. In both cases the extraction pro- 

 cess has been conducted, at the ordinary temperature, with almost similar 

 proportions of acid to soil, the acid used being, in Method I., approxi- 

 mately of 23J per cent, strength (M15 sp. gr.) and 25 per cent. (1'126 

 sp. gr.) in Method II. : in fact this, and the sifting of the soil through a 

 3 m.m. instead of a J m.m. sieve are the only apparent reasons for the 

 higher proportions both of potash and lime obtained from the soils by the 

 latter method. 



In the Witteberg soils of the Albany Division, for instance, an 

 average percentage of '1 of lime was obtained by Method II. Only in 

 soils of fairly good fertility has this proportion been yielded when the 

 first method was made use of. In the sandstone soils of Humansdorp, 

 too, such a result was noticed. 



The high results obtained in some cases by Method II. are noticeable 

 in the soils from the Bredasdorp Division, practically all of which were 

 examined by this method (the single exception analysed by Method I. 

 yielding lower results) and in those from the Caledon Division, although 

 it is to be observed that even by Method II. the Caledon soils show up 



